PROJECT 1. Sequential Multiple Assignment /Randomized Treatment (SMART) for BD: C Bowden, V Singh, P Thompson, C Martinez, J Calabrese, K Gao, M Quinones, J Mintz, M Tohen. A . What we propose to do. This open methods advancement study will randomize BD patients with clinically significant symptoms to treatment with one of two mood stabilizers (MS), lithium [LI] or divalproex [DV]. Those who develop protocol defined depression will then be randomized to a MS alone, MS + quetiapine [QT] or MS + lamotrigine [LM]. A SMART strategy employs a rule for adding new treatments based on each patient's current illness state and response during the trial, mimicking the adaptive nature of treatment selection which occurs in clinical settings, but in a controlled way which allows application of causal inference. By using early indices of response to dynamically alter treatment decisions to improve outcome, SMART eliminates unmeasured confounders associated with treatment decisions that are not randomized, as occurs in data mining exercises and in other non-randomized decisions in studies which randomize one variable at baseline. This sequential adaptive design represents a methodological innovation in bipolar trial history which will have particular implications for effectiveness studies.